brown@cyclps.DEC (Doug Brown) (05/02/84)
I've been pondering on a project that I was thinking of building
and would like to have your opinions on its viability.
What I have in mind is a low speed local area network device for device
control and communications. Below are the proposed characteristics:
max speed 19.2 K
Max nodes 255
max area 16
max local objects 16
max message size 255 bytes
max local devices 8
price ~$250 for controller
IT would be a single board that is highly configurable. Consisting of
a line modulator (coax, modem, twisted pair depending on speed), a dedicated
processor with rom control program and a device interface buss.
The main use would be for hooking up a series of devices in a small local
area with one controlling computer. Scenarios could be a small business,
house hold or manufacturing plant.
Devices for the controller would be swithes, serial lines, a/d, d/a.
Custom user programs up to 4K in size can be accomodated.
Hooking the whole thing up would be by tagged modules (a simpleton with a
screw driver should be able to do it for the very low speeds).
A simple software interface for custom devices.
What do you think? Would it be worth while?
.--. ... decvax!decwrl!rhea!cyclps!brown
|db| From the depths of DEC land.
`--' 2500+ nodes and counting...rpw3@fortune.UUCP (05/04/84)
#R:decwrl:-758300:fortune:5900017:000:259
fortune!rpw3 May 4 01:42:00 1984
Re: 19,200 b/s LAN
Try RS-422 (the versions of the chips with Tri-State outputs).
Rob Warnock
UUCP: {ihnp4,ucbvax!amd70,hpda,harpo,sri-unix,allegra}!fortune!rpw3
DDD: (415)595-8444
USPS: Fortune Systems Corp, 101 Twin Dolphin Drive, Redwood City, CA 94065mikez@hammer.UUCP (05/04/84)
Yes, a board with the suggest specs would be worthwhile. For about six months. <Major silicon stamper> is sampling out a low-speed version of their chip set that has all the Ethernet/IEEE 802 stuff and runs on twisted pair, for ~$25 a chip set. I expect that that's where the low speed network action will be.
mats@dual.UUCP (Mats Wichmann) (05/06/84)
Whether or not the suggested board is useful depends on its' intended
scope. I doubt whether the computer industry as a whole is going to
embrace the idea, but it might well find a niche somewhere. After all,
a LOT of people are using Corvus' rather awful OMNINET (that was a
personal opinion, folks - and don't bother flaming), and even more
would never consider touching it (like me). Even with its' rather
limited distribution, it would have to be called a successful
project. I think there is alway going to be room for a well-done,
low-cost LAN scheme.
As to the comment that the `scaled-down-ethernet' chips will be the
way people are going, this is opinion only. My opinion is that that
is NOT the way very many people are going to go. Specifically, I have
already seen the industry go IBM-compatible-crazy several times; I
would be very surprised if much of the development in the LAN area
during the next year or so is not in the IBM-syle net, whenever the
details of that become clear. And then there's always AT&T to consider...
One would *expect* them to try to emerge as a leader in networking,
since this is really in their line of business. Once again, opinions
only.
Anybody else have thoughts on this?
Mats Wichmann
Dual Systems Corp.
...{ucbvax,amd70,ihnp4,cbosgd,decwrl,fortune}!dual!mats